Nasal Dysbiosis in Cutaneous T-Cell Lymphoma Is Characterized by Shifts in Relative Abundances of Non-Staphylococcus Bacteria.

The nasal microbiome of patients with cutaneous T-cell lymphoma (CTCL) remains unexplored despite growing evidence connecting nasal bacteria to skin health and disease. Nasal swabs from 45 patients with CTCL (40 with mycosis fungoides, 5 with Sézary syndrome) and 20 healthy controls from the same geographical region (Chicago Metropolitan Area, Chicago, IL) were analyzed using sequencing of 16S ribosomal RNA and tuf2 gene amplicons. Nasal α-diversity did not differ between mycosis fungoides/Sézary syndrome and healthy controls (Shannon index, genus level, P = 0.201), but distinct microbial communities were identified at the class (R2 = 0.104, P = 0.023) and order (R2 = 0.0904, P = 0.038) levels. Increased relative abundance of the genera Catenococcus, Vibrio, Roseomonas, Acinetobacter, and unclassified Clostridiales was associated with increased skin disease burden (P < 0.005, q < 0.05). Performed to accurately resolve nasal Staphylococcus at the species level, tuf2 gene amplicon sequencing revealed no significant differences between mycosis fungoides/Sézary syndrome and healthy controls. Although S. aureus has been shown to worsen CTCL through its toxins, no increase in the relative abundance of this taxon was observed in nasal samples. Despite the lack of differences in Staphylococcus, the CTCL nasal microbiome was characterized by shifts in numerous other bacterial taxa. These data add to our understanding of the greater CTCL microbiome and provide context for comprehending nasal-skin and host‒tumor‒microbial relationships.

Introduction

Cutaneous T-cell lymphoma (CTCL) comprises a heterogeneous group of T-lymphocyte malignancies that infiltrate the skin. Patients with advanced, progressive disease often suffer from profound immune dysregulation and recurrent skin infections. Previous research suggests that the microbiome may influence CTCL pathogenesis, flares, and progression (Harkins et al., 2021; Lindahl et al., 2019; Willerslev-Olsen et al., 2013). Moreover, distinct microbe-precipitated metabolic and immunologic pathways have been linked to the pathobiology of atopic dermatitis (Nørreslet et al., 2020; Paller et al., 2019), psoriasis (Hidalgo-Cantabrana et al., 2019), hidradenitis suppurativa (McCarthy et al., 2022), and various malignancies (Goodman and Gardner, 2018)—conditions similarly known to be associated with immune dysregulation.

The ecosystem encompassing the nares may be a principal reservoir for self-contamination through nose-to-skin bacterial spread or vice versa. The importance of the nasal microbiome is further emphasized by recent literature suggesting that altered nasal bacterial diversity is associated with gut and skin dysbiosis in hidradenitis suppurativa (McCarthy et al., 2022). Although early culture-based studies have suggested that higher rates of Staphylococcus aureus skin and nasal colonization occur in patients with CTCL (Nguyen et al., 2008; Talpur et al., 2008), the complete nasal microbiome in CTCL has yet to be described. Although the CTCL skin microbiota is currently being investigated (Harkins et al., 2021; Salava et al., 2020), its nasal microbial profile is a missing piece of data because CTCL dysbiosis likely extends beyond the skin.

To better understand the CTCL nasal microbiome, we conducted a cross-sectional analysis of the nasal microbiota present in patients with CTCL and healthy controls (HCs) using 16S ribosomal RNA (rRNA) gene sequencing and further determined staphylococcal species relative abundances using tuf2 gene amplicon sequencing.

Results

Patient characteristics

A total of 45 patients comprised the patient group with CTCL, of which 40 had been diagnosed with mycosis fungoides (MF), and 5 had been diagnosed with Sézary syndrome (SS) (Table 1 and Supplementary Table S1). All patients and HCs were from the same geographical region (Chicago Metropolitan Area, Chicago, IL) to control for environmental influences on the microbiome (Rothschild et al., 2018). Four HC‒CTCL pairs sharing a home were selected for even closer matching. To avoid bias in sample collection, manipulation, and analysis, we concurrently enrolled patients and controls rather than rely on publicly available human microbiome data. There was no significant difference in age, sex, race/ethnicity, or phototype between the two groups (Table 1).

Table 1

Characteristics of Patients (n = 45) and Healthy Controls (n = 20)

Characteristics	Patients	Controls	P-Value
n	45	20
Sex1			0.27972
Male	29 (64.4)	10 (50.0)
Female	16 (35.6)	10 (50.0)
Age (y)3	62.7 (17.5–83.4)	54.5 (24.4–79.1)	0.13932
Race/Ethnicity1			0.90412
Asian	2 (4.4)	3 (15.0)
Black	5 (11.1)	0 (0.0)
White	30 (68.9)	15 (75.0)
White/Hispanic	6 (13.3)	1 (5.0)
Other/Hispanic	1 (2.2)	1 (5.0)
Phototype1			0.23982
Light (FST I‒III)	45 (100.0)	19 (95.0)
Dark (FST IV‒VI)	0 (0.0)	1 (5.0)
Comorbidities1
HTN	20 (44.4)	6 (30.0)	0.4114
DLP	24 (53.3)	7 (35.0)	0.1924
GERD	11 (24.4)	6 (30.0)	0.7614
Diagnosis Subtype1
MF	40 (88.9)	—
SS	5 (11.1)	—
Clinical stage1
Early (IA‒IIA)	26 (57.8)	—
Advanced (IIB‒IVB)	19 (42.2)	—
Disease duration (y)3	3.1 (0.2–30.0)	—
mSWAT3	22 (3–100)	—

Abbreviations: DLP, dyslipidemia; FST, Fitzpatrick skin phototype; GERD, gastroesophageal reflux; HTN, hypertension; MF, mycosis fungoides; mSWAT, modified Severity Weighted Assessment Tool; SS, Sézary syndrome.

Data are presented as n (%).

Data were analyzed with two-tailed t-test.

Data are presented as median (range).

Data were analyzed with Fisher’s exact test.

Supplementary Table S1

Detailed Demographic Characteristics of Patients with MF/SS (n = 45) and Healthy Controls (n = 20)

Gender	Age (y)	Race/Ethnicity	Diagnosis	Stage	mSWAT	Comorbidities	Skin-Directed Therapy	Systemic Therapy
Healthy controls (n = 20)
F	34	Asian	—	—	—	None	—	—
M	25	Asian	—	—	—	None	—	—
M	26	White	—	—	—	None	—	—
M	37	White	—	—	—	None	—	—
M	24	Asian	—	—	—	None	—	—
F	27	White/Hispanic	—	—	—	None	—	—
F	60	White	—	—	—	GERD, DLP, hypothyroidism, migraines, rheumatoid arthritis	—	—
F	68	White	—	—	—	DLP	—	—
M	76	White	—	—	—	GERD, DLP, HTN, T2DM	—	—
F	74	White	—	—	—	Asthma, CM, CKD (stage III), GERD, DLP, HTN, hypothyroidism, T2DM	—	—
F	54	White	—	—	—	None	—	—
M	55	White	—	—	—	None	—	—
F	65	White	—	—	—	GERD, DLP, T2DM	—	—
F	74	White	—	—	—	GERD, DLP, HTN	—	—
M	38	Other/Hispanic	—	—	—	GERD, HTN	—	—
M	79	White	—	—	—	HTN	—	—
F	54	White	—	—	—	None	—	—
F	58	White	—	—	—	DLP, HTN, hypothyroidism, T2DM	—	—
M	66	White	—	—	—	None	—	—
M	47	White	—	—	—	None	—	—
Patient with MF/SS (n = 45)
M	71	White	FMF	IIIB	35	Bladder cancer, GERD,DLP, T2DM	TCS	Acitretin
M	65	White	FMF	IIIA	90	GERD	TCS	—
F	18	White/Hispanic	FMF	IB	26	Asthma, GERD	TCS, NBUVB	IFN-α-2b
M	56	White/Hispanic	FMF	IB	16	Anemia, CKD, GERD, glaucoma, DLP, HTN, T2DM, vertebral osteomyelitis	TCS, TCI, NBUVB	—
F	72	White	FMF	IB	5	Diverticulitis, GERD, DLP, hypothyroidism, infiltrating ductal carcinoma	TCS	IFN-α-2b, bexarotene
M	36	White/Hispanic	FMF	IB	3	DLP, T2DM	TCS	—
F	74	White	FMF	IIB	18	DLP, HTN	TCS, TCI	Acitretin
F	35	Other/Hispanic	FMF	IB	28	None	—	—
F	63	Black	FMF	IIB	13	DLP, HTN, hypothyroidism, obesity	TCS, NBUVB	—
M	67	White	FMF/SMF	IB	13	DLP, HTN	TCS, NBUVB, XRT	IFN-α-2b, acitretin
M	55	White	FMF/SMF	IB	3	None	TCS, NBUVB	—
M	69	White	CD4+ MF	IIB	7	CAD, insomnia, obesity, overactive bladder	TCS	—
M	49	Asian	CD4+ MF	IB	60	Cataracts, ED, mitral regurgitation, obesity, T2DM	—	—
M	37	White/Hispanic	CD4+ MF	IIB	65	None	TCS, NBUVB	IFN-α-2b
M	47	White	CD4+ MF	IB	4	GERD, obesity	TCS	—
M	58	White	CD4+ MF	IIA	22	Autoimmune hemolytic anemia, anaplastic large cell lymphoma, asthma, HTN, hyperthyroidism, obesity, OSA	TCS	Acitretin
M	65	White	CD4+ MF	IA	13	GERD, HTN, DLP, OSA, T2DM	TCS	—
F	37	White/Hispanic	CD4+ MF	IB	25	HTN, hypothyroidism, OSA	TCS, NBUVB	Bexarotene, acitretin
F	72	Black	CD4+ MF	IIIA	100	DLP, HTN, T2DM	TCS	—
M	69	White	CD4+ MF	IIIA	14	AF, BPH, CAD, DLP, T2DM	TCS	—
M	67	White	CD4+ MF	IIB	4.5	CRC, DLP	TCS, Imiquimod	Acitretin
M	52	White	CD4+ MF	IB	13	Allergic rhinitis	TCS	—
M	81	White	CD4+ MF	IB	10	BPH, DLP, PAF, prostate cancer	TCS	Bexarotene
F	74	White	CD4+ MF	IA	5	Hypothyroidism	TCS	—
M	65	White	CD4+ MF	IA	5	DLP, GERD, HTN, obesity, T2DM	TCS	—
M	61	White	CD4+ MF	IA	6	HTN	TCS	—
F	26	White	CD4+ MF	IB	68	Migraine	—	—
M	63	White	CD4+ MF	IIB	6	Hypothyroidism	TCS	—
F	72	Black	CD4+ MF	IIIA	85	Glaucoma, DLP, HTN	TCS	—
F	61	White	CD4+ MF	IA	2	GERD, endometriosis, osteoporosis, primary hyperparathyroidism	—	—
M	74	White	CD4+ MF	IA	22	Osteoarthritis, BPH, CKD, GERD, glaucoma, HTN, ulnar neuropathy	TCS	Methotrexate, bexarotene
M	46	Black	CD4+ MF	IIB	60	Asthma, bronchitis	TCS	—
M	62	White	CD4+ MF	IB	23	DLP	—	—
F	55	White	CD4+ MF	IB	25	Asthma, COPD, GERD, DLP, HTN, PAD, PFO	—	—
M	68	White	CD4+ MF	IIB	50	DLP, HTN	—	—
M	55	White	CD4+ MF	IB	10	DLP, HTN, obesity	—	—
M	72	White	CD4+ MF	IIIB	95	Atrial flutter, anxiety, cervical stenosis, CKD, ED, DLP, HTN, lumbar disc herniation, OSA, PAF, prostate cancer, tachycardia-induced CM	TCS	Brentuximab vedotin
F	58	Asian	CD4+ MF	IB	14	Breast cancer, DLP, HTN, multinodular goiter, T2DM	TCS	—
M	29	White/Hispanic	CD8+ MF	IB	24	None	TCS	—
F	46	White	CD8+ MF	IA	15	Allergic rhinitis, fibroids, obesity	—	—
M	66	White	SS	IV	10	AF, dysphagia, DLBCL, DLP, HSV keratitis, hypothyroidism	TCS	—
F	56	White	SS	IV	24	DLP, HTN, hypothyroidism	TCS	—
M	83	White	SS	IV	100	CAD, HTN	TCS	Methotrexate
F	66	White	SS	IV	28	DLP, hypothyroidism	TCS	—
M	67	Black	SS	IV	32	Atrial flutter, CAD, ED, HFrEF, DLP, HTN, obesity, T2DM	TCS	—

Abbreviations: AF, atrial fibrillation; BPH, benign prostatic hyperplasia; CAD, coronary artery disease; CM, cardiomyopathy; CKD, chronic kidney disease; COPD, chronic obstructive pulmonary disease; CRC, colorectal carcinoma; DLBCL, diffuse large B-cell lymphoma; DLP, dyslipidemia; ED, erectile dysfunction; F, female; FMF, folliculotropic mycosis fungoides; GERD, gastroesophageal reflux disease; HFrEF, heart failure with reduced ejection fraction; HSV, herpes simplex virus; HTN, hypertension; LCT, large cell transformation; M, male; MF, mycosis fungoides; mSWAT, modified Severity Weighted Assessment Tool; NBUVB, narrowband UVB; OSA, obstructive sleep apnea; PAD, peripheral artery disease; PAF, paroxysmal atrial fibrillation; PFO, patent foramen ovale; SMF, syringotropic mycosis fungoides; SS, Sézary syndrome; T2DM, type 2 diabetes mellitus; TCI, topical calcineurin inhibitor; TCS, topical corticosteroids; XRT, localized radiotherapy.

A total of 26 patients had the early-stage disease (stages IA‒IIA; 57.8%), and 19 had the advanced-stage disease (stages IIB‒IVB; 42.2%); stage IB was the most common overall (n = 18, 40.0%). The median modified Severity-Weight Assessment Tool (mSWAT) score was 22 (range = 3‒100). The most common comorbidities reported were dyslipidemia (MF/SS = 53.3%, HCs = 35.0%), hypertension (MF/SS = 44.4%, HCs = 30.0%), and gastroesophageal reflux (MF/SS = 24.4%, HCs = 30.0%). There were no significant differences in comorbidities between the two groups for these three conditions (Fisher’s exact test: dyslipidemia P = 0.192; hypertension P = 0.411; and gastroesophageal reflux P = 0.761) or for any other comorbidity.

Differences in nasal microbiota between patients and controls

The 16S rRNA gene amplicon sequence data identified a total of 720 genera, 285 families, 139 orders, 67 classes, and 28 phyla. Swab, reagent, and PCR controls were negative for any significant contamination. The most abundant phyla in both groups were the three most frequently encountered in the human nares: Proteobacteria, Actinobacteria, and Firmicutes. At the genus level, there was no significant difference in biodiversity between MF/SS and HC samples as assessed by Shannon diversity index (P = 0.201) (Figure 1a). Notably, β-diversity revealed a small but globally significant difference in the microbial community structure between patients and controls on the basis of Adonis/permutational ANOVA (R2 = 0.104, P = 0.023 for class level; R2 = 0.0904, P = 0.038 for order level) (Figure 1b and c). The most abundant genera in both groups were Corynebacterium and Staphylococcus (Figure 1d).

Figure 1

Distinct nasal bacterial communities were identified in patients with MF/SS versus in HCs. (a) α-Diversity was not significantly different between HCs and patients with MF/SS at the genus level (Shannon diversity index, P = 0.201). MDS plots using the Bray‒Curtis dissimilarity index of β-diversity analyses show significant differential clustering of HCs and patients with MF/SS at the taxonomic levels of (b) class (Adonis/PERMANOVA R2 = 0.104, P = 0.023) and (c) order (R2 = 0.0904, P = 0.038). (d) Relative abundance (%) of the 20 most abundant genera in nasal samples of HCs and patients with MF/SS (left, individual subjects; right, mean relative abundances per group [HC, MF/SS]). HC, healthy control; MDS, multidimensional scaling; MF, mycosis fungoides; PERMANOVA, permutational ANOVA; SS, Sézary syndrome.

Specific taxa contributing to the distinct nasal microbiota of patients with MF/SS were then investigated. Several genera were significantly higher in patients than in HCs (q < 0.05): Roseomonas, Catenococcus, Vibrio, Marinobacter, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, Acinetobacter, Alishewanella, Paracoccus, unclassified Clostridiales and unclassified Clostridiales family XIII, Atopobium, and Dietzia (Table 2 and Figure 2). Meanwhile, Lachnospiraceae NK4A136 group was reduced in patient samples (q < 0.05). Regression analyses revealed a positive association between the relative abundance of Catenococcus, Vibrio, Roseomonas, unclassified Clostridiales, and Acinetobacter genera and increased mSWAT, an indicator of skin disease burden; reduced Lachnospiraceae NK4A136 group relative abundance was associated with higher mSWAT scores (Figure 3). One-way ANOVA revealed significant differences in the mean relative abundance of Alishewanella, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, Marinobacter, and Vibrio between HCs and patients as grouped by low versus high mSWAT and early versus advanced disease (Supplementary Table S2, Supplementary Table S3, Supplementary Table S4, Supplementary Table S5). Sidak test for pairwise comparisons showed that the mean relative abundance of Vibrio was significantly different between HCs and patients with low (P = 0.034) and high (P = 0.004) mSWAT and between HCs and patients with early-stage (P = 0.011) and advanced-stage (P = 0.007) disease.

Table 2

Differential Taxonomic Analysis Shows Unique Microbial Signatures at the Genus Level in Nasal Samples from Patients with MF/SS Versus HCs

	Genus	HC/Patient LogFC	P-Value	q-Value1
Reduced abundance in MF/SS	Lachnospiraceae NK4A136 group	0.51	0.005	0.04
	Ruminococcus	0.12	0.02	0.14
	Ruminoclostridium	0.48	0.03	0.15
Enriched abundance in MF/SS	Catenococcus	‒1.41	<0.001	0.001
	Alishewanella	‒1.07	<0.001	0.001
	Vibrio	‒1.96	<0.001	0.001
	Unclassified Clostridiales family XIII	‒0.35	<0.001	0.001
	Roseomonas	‒0.89	<0.001	0.001
	Unclassified Clostridiales	‒0.62	<0.001	0.001
	Paracoccus	‒1.36	<0.001	0.001
	Marinobacter	‒1.07	<0.001	0.001
	Atopobium	‒0.38	<0.001	<0.005
	Dietzia	‒0.62	<0.001	<0.005
	Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium	‒1.32	<0.001	<0.01
	Acinetobacter	‒0.99	0.005	0.04
	Unclassified bacteria	‒0.59	0.01	0.09
	Christensenellaceae R-7 group	‒0.04	0.01	0.09
	Cutibacterium	‒0.21	0.02	0.11
	Escherichia/Shigella	‒1.04	0.02	0.13
	Neisseria	‒0.81	0.03	0.14
	Pseudoalteromonas	‒1.7	0.03	0.14
	Subdoligranulum	‒0.53	0.04	0.15
	Veillonella	‒0.93	0.04	0.17
	Actinomyces	‒1.18	0.04	0.17
	Unclassified Gammaproteobacteria	‒0.64	0.05	0.17

Abbreviations: FC, fold change; HC, healthy control; MF, mycosis fungoides; SS, Sézary syndrome.

The q-value is the FDR-adjusted P-value (Benjamini and Hochberg, 1995).

Figure 2

Changes in the abundance of specific bacterial genera present in the nares of patients with MF/SS compared with those in the nares of the HCs. Dot plots illustrate the relative sequence abundance (%) of genera that are (a) significantly enriched and (b) significantly reduced in patients with MF/SS versus in HCs. Mean relative abundances are indicated by black horizontal bars. Significance is determined by q ≤ 0.05; the q-value is the FDR-adjusted P-value (Benjamini and Hochberg, 1995). FDR, false discovery rate; HC, healthy control; MF, mycosis fungoides; SS, Sézary syndrome.

Figure 3

Relationship between the relative abundances of significantly enriched and depleted genera and skin disease burden in patients with MF/SS. The relative abundance (%) of each genus is plotted versus mSWAT score (an indicator of skin disease burden) with line of best fit. (a) Increased mSWAT score was associated with an increased relative abundance of several genera that were enriched in patients with MF/SS: Catenococcus, Vibrio, Roseomonas, unclassified Clostridiales, and Acinetobacter. (b) Lower relative abundances were associated with increased mSWAT scores for the remaining enriched genera: Paracoccus, Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium, Alishewanella, Marinobacter, Dietzia, unclassified Clostridiales family XIII, and Atopobium. (c) Regression analysis of Lachnospiraceae NK4A136 group (reduced in patients with MF/SS compared with that in the HCs) revealed that lower relative abundances were associated with higher mSWAT scores. HC, healthy control; MF, mycosis fungoides; mSWAT, modified Severity-Weight Assessment Tool; SS, Sézary syndrome.

Supplementary Table S2

Mean Relative Abundances (%) of Genera Identified on Differential Analysis Comparing Bacterial Communities in the Samples of HCs Versus in Patients with MF/SS—One-Way ANOVA: HCs Versus MF/SS, Organized by mSWAT Score

Genus	HCs	Patients with MF/SS		P-Value
		mSWAT < 10	mSWAT ≥ 10
Acinetobacter	1.123	1.132	1.213	0.386
Alishewanella	0.729	0.884	0.807	0.006
Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium	0.729	0.869	0.850	0.005
Atopobium	0.693	0.748	0.700	0.405
Catenococcus	0.984	1.079	1.115	0.087
Dietzia	0.701	0.788	0.727	0.329
Lachnospiraceae NK4A136 group	0.835	0.786	0.740	0.269
Marinobacter	0.761	0.882	0.849	0.042
Paracoccus	0.699	0.887	0.807	0.054
Roseomonas	0.707	0.742	0.792	0.303
Unclassified Clostridiales	0.693	0.731	0.736	0.378
Unclassified Clostridiales family XIII	0.698	0.713	0.708	0.950
Vibrio	0.912	1.111	1.104	0.003

Abbreviations: HC, healthy control; MF, mycosis fungoides; mSWAT, modified Severity Weighted Assessment Tool; SS, Sézary syndrome.

Supplementary Table S3

Mean Relative Abundances (%) of Genera Identified on Differential Analysis Comparing Bacterial Communities in the Samples of HCs Versus in Patients with MF/SS—Sidak Method for Pairwise Comparisons: HCs Versus MF/SS, Organized by mSWAT Score

Group 1	Group 2	Mean Difference	P-Value
Alishewanella
HC	MF/SS, mSWAT < 10	0.155	0.005
HC	MF/SS, mSWAT ≥ 10	0.078	0.096
MF/SS, mSWAT < 10	MF/SS, mSWAT ≥ 10	‒0.077	0.232
Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium
HC	MF/SS, mSWAT < 10	0.140	0.028
HC	MF/SS, mSWAT ≥ 10	0.121	0.009
MF/SS, mSWAT < 10	MF/SS, mSWAT ≥ 10	‒0.019	0.971
Marinobacter
HC	MF/SS, mSWAT < 10	0.120	0.082
HC	MF/SS, mSWAT ≥ 10	0.088	0.095
MF/SS, mSWAT < 10	MF/SS, mSWAT ≥ 10	‒0.03	0.885
Vibrio
HC	MF/SS, mSWAT < 10	0.200	0.034
HC	MF/SS, mSWAT ≥ 10	0.192	0.004
MF/SS, mSWAT < 10	MF/SS, mSWAT ≥ 10	‒0.008	0.999

Abbreviations: HC, healthy control; MF, mycosis fungoides; mSWAT, modified Severity Weighted Assessment Tool; SS, Sézary syndrome.

Supplementary Table S4

Mean Relative Abundances (%) of Genera Identified on Differential Analysis Comparing Bacterial Communities in the Samples of HCs Versus in Patients with MF/SS—One-Way ANOVA: HC Versus MF/SS, Organized by Clinical Stage

Genus	HC	Pateints with MF/SS		P-Value
		IA‒IIA	IIB‒IVB
Acinetobacter	1.123	1.181	1.210	0.551
Alishewanella	0.729	0.843	0.803	0.016
Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium	0.729	0.881	0.818	0.002
Atopobium	0.693	0.692	0.739	0.332
Catenococcus	0.984	1.157	1.036	0.015
Dietzia	0.701	0.736	0.750	0.600
Lachnospiraceae NK4A136 group	0.835	0.707	0.813	0.077
Marinobacter	0.761	0.868	0.842	0.044
Paracoccus	0.699	0.858	0.784	0.049
Roseomonas	0.707	0.714	0.869	0.011
Unclassified Clostridiales	0.693	0.720	0.755	0.220
Unclassified Clostridiales family XIII	0.698	0.739	0.669	0.242
Vibrio	0.912	1.096	1.119	0.003

Abbreviations: HC, healthy control; MF, mycosis fungoides; SS, Sézary syndrome.

Supplementary Table S5

Mean Relative Abundances (%) of Genera Identified on Differential Analysis Comparing Bacterial Communities in the Samples of HCs Versus in Patients with MF/SS—Sidak Method for Pairwise Comparisons: HC Versus MF/SS, Organized by Clinical Stage

Group 1	Group 2	Mean Difference	P-Value
Alishewanella
HC	MF/SS, IA‒IIA	0.114	0.013
HC	MF/SS, IIB‒IVB	0.074	0.220
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	‒0.040	0.663
Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium
HC	MF/SS, IA‒IIA	0.152	0.001
HC	MF/SS, IIB‒IVB	0.089	0.130
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	‒0.063	0.350
Catenococcus
HC	MF/SS, IA‒IIA	0.173	0.016
HC	MF/SS, IIB‒IVB	0.053	0.800
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	‒0.120	0.148
Marinobacter
HC	MF/SS, IA‒IIA	0.106	0.044
HC	MF/SS, IIB‒IVB	0.081	0.225
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	‒0.026	0.912
Paracoccus
HC	MF/SS, IA‒IIA	0.159	0.043
HC	MF/SS, IIB‒IVB	0.085	0.512
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	‒0.744	0.581
Roseomonas
HC	MF/SS, IA‒IIA	0.159	0.043
HC	MF/SS, IIB‒IVB	0.085	0.512
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	‒0.744	0.581
Vibrio
HC	MF/SS, IA‒IIA	0.184	0.011
HC	MF/SS, IIB‒IVB	0.207	0.007
MF/SS, IA‒IIB	MF/SS, IIB‒IVB	0.024	0.974

Abbreviations: HC, healthy control; MF, mycosis fungoides; SS, Sézary syndrome.

Given the known role of S. aureus in worsening CTCL through its toxins (Fujii, 2022), we next examined whether the relative abundances of Staphylococcus species differed between patients with MF/SS and HCs. We performed targeted sequencing of the bacterial tuf gene, which provides accurate species-level resolution of Staphylococcus communities (Ahle et al., 2021). S. epidermidis and S. aureus were the most abundant staphylococcal species in both groups: these species comprised 56.8% and 23.8% of all staphylococcal species for patients with MF/SS and 52.7% and 14.5% for HCs, respectively (Supplementary Figure S1). There was no statistically significant difference between the relative abundance of any Staphylococcus species (including S. aureus and S. epidermidis) between patients with MF/SS and HCs (Supplementary Table S6).

Supplementary Figure S1

Relative abundance of staphylococcal species in patients with MF/SS compared with those in HCs. (a) Relative abundance (%) of Staphylococcus species present in the nasal samples of HCs and patients by individual subjects. (b) Mean relative abundances (%) per study group. HC, healthy control; MF, mycosis fungoides; SS, Sézary syndrome.

Supplementary Table S6

Differential Analysis for Staphylococcus Species Reveals No Statistically Significant Differences between the Nasal Microbiota of Patients with MF/SS and HCs

	Staphylococcus Species	Patient/HC LogFC	P-Value	q-Value1
Reduced abundance in MF/SS	S. lugdunensis	‒2.54	0.14	0.43
	S. hominis	‒1.92	0.16	0.43
	S. capitis	‒0.29	0.73	0.82
Enriched abundance in MF/SS	Unclassified Staphylococcus	1.48	0.22	0.43
	S. aureus	0.72	0.82	0.82
	S. epidermidis	0.19	0.74	0.82

Abbreviations: FC, fold change; FDR, false discovery rate; HC, healthy control; MF, mycosis fungoides; SS, Sézary syndrome.

The q-value is the FDR-adjusted P-value (Benjamini and Hochberg, 1995).

Discussion

Our results show that the nasal microbiomes of patients with MF/SS and HCs are different. These data add to the existing body of knowledge that supports the importance of the nasal microbiome in skin disease (McCarthy et al., 2022; Olesen et al., 2021; Totté et al., 2019). Nasal microbiota are already known to be important in atopic dermatitis, in which increased relative abundance of nasal Staphylococcus and Moraxella and decreased Dolosigranulum are associated with disease severity (Totté et al., 2019), and increased nasal S. hominis is linked to skin S. hominis abundance and disease improvement (Olesen et al., 2021). The nares of patients with hidradenitis suppurativa are characterized by enriched Proteus communities and reduced Corynebacterium (McCarthy et al., 2022). In addition, loss of nasal Proteobacteria has been associated with skin and soft tissue infections (Johnson et al., 2015), and nasal S. aureus colonization has been implicated in disease activity in various inflammatory skin conditions, including CTCL (Ng et al., 2017; Nørreslet et al., 2020; Talpur et al., 2008). The nasal microbiome could also feasibly serve as a source for bacterial recolonization of the skin after systemic antibiotic treatment (Lindahl et al., 2021), if not also increase the risk of recurrent infections.

Although the exact mechanisms for how the nasal microbiome influences CTCL pathogenesis and vice versa are unclear, the data included in this study provide greater context from which ongoing CTCL skin microbiome research can be understood. We found that enrichment of the genera Vibrio, Roseomonas, and Acinetobacter and depletion of Paracoccus are associated with increased skin severity. From the literature, we know that Vibrio, Roseomonas, and Acinetobacter bacteria are important in causing necrotizing fasciitis, aggravating atopic dermatitis, and instigating skin and soft tissue infections, respectively (Table 3 and Supplementary Discussion) (Cerqueira and Peleg, 2011; Janda et al., 1988; Myles et al., 2018); however, their role as nasal bacteria requires further study because there remains an extreme paucity of knowledge on the biological relationships shared by non‒Staphylococcus species in the nasal microbiome in healthy and disease states.

Table 3

Summary of Human Disease Associations of Significantly Enriched/Reduced Genera Found in the Anterior Nares of Patients with MF/SS

Genus	Associations with Human Disease	Associations with Human Cutaneous Disease
Acinetobacter	Hospital- and community-acquired pneumonia, invasive bloodstream infections, urinary tract infections, hospital-acquired meningitis, osteomyelitis, pericarditis	Skin and soft tissue infections
Alishewanella	None available	None available
Allorhizobium-Neorhizobium-Pararhizobium-Rhizobium	None available	None available
Atopobium	Bacteremia, dental infections, bacterial vaginosis	Bacteremia in the setting of Fournier’s gangrene
Catenococcus	None available	None available
Dietzia	Bacteremia, prosthetic hip infection, pacemaker infection, pleural fluid isolate	Confluent and reticulated papillomatosis
Lachnospiraceae NK4A136 group	Decreased abundance in the gut microbiome after Trichinella spiralis infection and in patients with dementia; biomarker for lean body habitus	None available
Marinobacter	None available	None available
Paracoccus	P. yeei: myocarditis, peritonitis, bacteremia	None available
Roseomonas	SepticemiaR. mucosa: catheter-related infections, dialysis and surgical wound infections, bacteremia	Skin and soft tissue infections, atopic dermatitis
Unclassified Clostridiales	Mediates allergic immune activity	Reduced in the gut microbiome of alopecia areata and pediatric atopic dermatitis
Unclassified Clostridiales family XIII	Mediates mood disorders	None available
Vibrio	Cholera, gastroenteritis, sepsis, less commonly otitis media, meningitis, peritonitis, and pneumonia	Necrotizing fasciitis

Abbreviations: MF, mycosis fungoides; SS, Sézary syndrome.

Importantly, our data showed that the nasal relative abundances of Staphylococcus species in patients with CTCL did not differ significantly from those of HCs. Although the pathogenic role of S. aureus toxins in CTCL has been well established over the years (Fujii, 2022), it had been unclear whether this translates to a higher relative abundance of S. aureus in the skin and/or nose. Previous culture-based studies have either failed to show a statistically significant difference in nasal S. aureus colonization rates between patients with CTCL and HCs (Nguyen et al., 2008) or did not include a matched comparison group (Talpur et al., 2008). Our nasal data, together with recent CTCL skin microbiome data (Harkins et al., 2021), suggest that the effects of S. aureus in CTCL may not translate to the actual increased relative abundance of S. aureus. Instead, it remains possible that S. aureus toxin production—and not relative abundance—differs between patients with CTCL and HCs. These differences between patients and controls may be mediated by shifts in the abundances of the other bacterial taxa. In fact, in atopic dermatitis, quorum sensing between bacterial species in the skin revealed that coagulase-negative staphylococci species produce autoinducing peptides that inhibit S. aureus phenol-soluble modulin α, a proinflammatory virulence factor capable of mediating epidermal injury (Williams et al., 2019).

Nasal dysbiosis carries intriguing insights for pathophysiology in a disease where advanced-stage patients often suffer from recurrent skin infections (Blaizot et al., 2018). Through the accurate characterization of the nasal microbial profiles associated with worse disease, we can conceivably intervene by altering the nasal microbiome through the decolonization of high-risk bacteria or reconstitution with bacteria associated with healthy individuals. The nasal microbiome may have the potential to serve as an additional and accessible biomarker for the determination of disease progression risk. Eventual matched patient skin and nasal microbiome analyses can further elucidate these relationships.

In this study, we establish that CTCL is characterized by nasal dysbiosis composed of shifts in specific non‒Staphylococcus taxa compared with that of healthy individuals. Because bacterial activity perpetuates CTCL disease progression and because infection is the most common cause of death in this patient population (Tsambiras, 2001; Willerslev-Olsen et al., 2013), attention to the nasal microbiome and its relationship with other microbial reservoirs is crucial to our understanding of the CTCL disease state and pathogenesis.

Materials and Methods

Participants

Ethical approval was obtained from the Northwestern University Institutional Review Board (STU00209226). Written informed consent, nasal samples, and personal data were obtained at the Northwestern University Cutaneous Lymphoma Clinic (Chicago, Illinois) between 2019 and 2021 in compliance with the Declaration of Helsinki. Each patient had clinically and biopsy-proven CTCL, as reviewed by an expert dermatopathologist (JG). At the time of sample collection, patients were receiving standard-of-care therapies, including skin-directed (n = 36, 80.0%) and select systemic (n = 13, 28.9%) treatments or were treatment naive (n = 9, 18.9%) (Supplementary Table S1). Subjects on any antibiotics within the preceding 4 weeks were excluded. Clinical staging and mSWAT were assessed by the study’s principal investigator (XAZ) at sample collection. The HC group (n = 20) was composed of age-matched volunteers without CTCL or other skin diseases from the same geographical region.

Sample collection and DNA extraction

Nasal samples were obtained through sterile swabs (FLOQSwabs, Copan Diagnostics, Murrieta, CA) with hands covered in sterile gloves. All specimens were placed immediately in sterile cryovials and promptly stored at ‒80°C until DNA extraction. Genomic DNA was extracted using a Maxwell 16 LEV Blood DNA Kit (Promega, Madison, WI) implemented on a Maxwell 16 Instrument, following the manufacturer’s instructions with minor modifications: a lysozyme incubation (10 ng/μl lysozyme; Thermo Fisher Scientific, Waltham, MA) for 30 minutes at 37°C and bead beating (40 seconds at 6 min/sec) using a FastPrep-24 System (MP Biomedicals, Irvine, CA). Homogenized samples were transferred to the Maxwell cartridges for final DNA purification.

16S rRNA gene amplicon sequencing

Genomic DNA was prepared for sequencing using a two-stage amplicon sequencing workflow, as described previously (Naqib et al., 2018), using primers targeting the V4 (fourth hypervariable) region of microbial 16S rRNA genes. The 515 forward modified and 806 reverse modified primers contained 5′ linker sequences compatible with access array primers for Illumina sequencers (Fluidigm, South San Francisco, CA) (Walters et al., 2015). PCRs were performed in a total volume of 10 μl using MyTaq HS 2X Mix (Meridian Bioscience, Cincinnati, OH) primers at 500 nM concentration and approximately 1,000 copies per reaction of a synthetic double-stranded DNA template (described below). Extraction blanks and PCR blanks were treated as independent samples and sequenced with unique barcodes. Thermocycling conditions were 95 °C for 5 minutes (initial denaturation), followed by 28 cycles of 95°C for 30 seconds, 55°C for 45 seconds, and 72°C for 30 seconds. Second-stage reactions contained 1 μl of PCR product and a unique primer pair of access array primers; thermocycling conditions consisted of 95°C for 5 minutes (initial denaturation), followed by 8 cycles of 95°C for 30 seconds, 60°C for 30 seconds, and 72°C for 30 seconds. Libraries were pooled and sequenced on an Illumina MiniSeq sequencer (Illumina, San Diego, CA) with 15% phiX spike-in and paired-end 2 × 153 base sequencing reads.

A synthetic double-stranded DNA spike-in was synthesized as a gBLOCK by Integrated DNA Technologies (Coralville, IA). The basis of the design was a 999 base pairs region of the 16S rRNA gene of Rhodanobacter denitrificans strain 2APBS1T (NC_020541) (Prakash et al., 2012). Portions of V1, V2, and V4 variable regions were replaced by eukaryotic mRNA sequences (Apostichopus japonicus Gapdh mRNA, HQ292612; and Strongylocentrotus intermedius Gapdh mRNA, KC775387). Primer sites were preserved, and the overall length in the base pair of the synthetic DNA did not differ from the equivalent R. denitrificans fragment. PCR amplicons generated from this synthetic DNA do not differ in size from bacterial amplicons and can only be identified and removed through postsequencing bioinformatics analysis. The sequence can be accessed through GenBank using the accession number OK324963.

tuf2 amplicon next-generation sequencing

Genomic DNA was PCR amplified with primers ACACTGACGACATGGTTCTACAACAGGCCGTGTTGAACGTG for CS1_tuf2 forward and TACGGTAGCAGAGACTTGGTCTACAGTACGTCCACCTTCACG for CS2_tuf2 reverse (Ahle et al., 2021, 2020) targeting the Staphylococcus tuf gene. Amplicons were generated using a two-stage PCR amplification protocol as previously described (Naqib et al., 2018). First-stage PCR amplifications were performed in 10 μl reactions in 96-well plates using MyTaq HS 2X mastermix (Meridian Bioscience). PCR conditions were 95°C for 5 minutes, followed by 28 cycles of 95°C for 30 seconds, 55°C for 30 seconds, and 72°C for 60 seconds. Second-stage reactions using access array primers were performed as described earlier. Samples were pooled, purified, and sequenced on an Illumina MiSeq with 10% phiX spike-in and paired-end 2 × 300 base sequencing reads (i.e., V3 chemistry). Library preparation, pooling, and sequencing were performed at the Genome Research Core within the Research Resources Center at the University of Illinois Chicago (Chicago, IL).

Sequence data processing (DADA2)

To check for contamination, control swab, PCR, and reagent/kit samples were performed. In total, 22 PCR and 32 extraction controls were analyzed, all of which yielded very low sequence counts (mean ± SD: 122.5 ± 61.4), far below the 5,000 counts per sample threshold needed for inclusion in data analyses.

16S rRNA gene amplicon reads were merged using PEAR, version 0.9.6 (Zhang et al., 2014), and trimmed using cutadapt, version 1.18, to remove ambiguous nucleotides and primer sequences on the basis of a quality threshold of P = 0.01 (Martin, 2011). Reads lacking the primer sequence and/or sequences <225 base pairs after merging and quality trimming were discarded. Chimeric sequences were identified and removed using the USEARCH algorithm with a comparison with Silva (version 132) reference sequence (Edgar, 2010; Glöckner et al., 2017). Amplicon sequence variants were identified using DADA2, version 1.18 (Callahan et al., 2016), and annotated taxonomically using the Naive Bayesian classifier included in DADA2 with the Silva (version 132) training set. Synthetic spike-in sequences were removed before proceeding with downstream bioinformatics analyses. Diversity analyses were performed in R using the vegan library, version 2.5-6 (Okansen et al., 2018). Biodiversity (α-diversity) was calculated using the Shannon index modeled with the sample covariates using a generalized linear model assuming Gaussian distribution. Bray‒Curtis indices were calculated to assess sample dissimilarity (β-diversity).

For the tuf2 next-generation sequencing, merged reads that lacked either primer sequence or were <400 base pairs were discarded. Chimeric sequences were identified and removed in a de novo fashion using USEARCH, version 8.1.1861 (Edgar, 2010). Amplicon sequence variants were identified using the protocol described earlier and taxonomically annotated using alignment from BLAST (blastn) with the RefSeq Prokaryotic Genomes reference, downloaded on 1 December 2021 (NCBI Resource Coordinators, 2017).

Differential analysis of microbial taxa

Differential analyses of taxa as compared with experimental covariates were performed using edgeR (version 3.28.1) on raw sequence counts (McCarthy et al., 2012). The 16S data were filtered to remove sequences of chloroplast, mitochondrial, or eukaryotic origin and taxa present in <30% of all samples and with <500 total sequence counts across all samples. The tuf2 next-generation sequencing data were filtered to retain only species belonging to the genus Staphylococcus and to remove taxa following the same parameters as mentioned earlier. Data were normalized as counts per million and fit using a negative binomial generalized linear model using experimental covariates.

Statistical analyses

Statistical analyses were performed in R and STATA SE. Significance of the α-diversity model (ANOVA) was tested using the F-test. Posthoc, pairwise analyses were performed using the Mann‒Whitney test (Wickham, 2009). The dissimilarity indices were tested for significance using Adonis/permutational ANOVA, and additional comparisons of the individual covariates were performed using analysis of similarities. Statistical tests for the differential analyses were performed using a likelihood ratio test. Adjusted P-values (q-values) were calculated using the Benjamini‒Hochberg false discovery rate correction (Benjamini and Hochberg, 1995). Significant taxa were determined on the basis of a false discovery rate threshold of 5.0% (0.05). Plots were generated using GraphPad Prism, version 9.2, (GraphPad Software, San Diego, CA) and the ggplot2 library in R (Wickham, 2009).

Data availability statement

Datasets related to this article can be found at https://dataview.ncbi.nlm.nih.gov/object/PRJNA768111?reviewer=pd94ec0d6iurp8k0gbs5evtjhj (National Center for Biotechnology Information Short Read Archive, accession number PRJNA768111).

ORCIDs

Madeline J. Hooper: http://orcid.org/0000-0003-1334-5342

Tessa M. LeWitt: http://orcid.org/0000-0001-8935-2165

Francesca L. Veon: http://orcid.org/0000-0002-0232-1671

Yanzhen Pang: http://orcid.org/0000-0003-1825-9603

George E. Chlipala: http://orcid.org/0000-0003-0203-3191

Leo Feferman: http://orcid.org/0000-0002-5821-3434

Stefan J. Green: http://orcid.org/0000-0003-2781-359X

Dagmar Sweeney: http://orcid.org/0000-0001-6320-8931

Katherine T. Bagnowski: http://orcid.org/0000-0002-7482-4454

Michael B. Burns: http://orcid.org/0000-0001-9791-4359

Patrick C. Seed: http://orcid.org/0000-0001-8998-8374

Joan Guitart: http://orcid.org/0000-0001-7635-9237

Xiaolong A. Zhou: http://orcid.org/0000-0002-6177-2472

Author Contributions

Conceptualization: XAZ, TML, FLV, MJH, JG; Data Curation: XAZ, TML, YP, KTB; Formal Analysis: GEC, LF, SJG, XAZ; Funding Acquisition: XAZ; Investigation: XAZ, TML, FLV, MJH, YP, KTB, DS; Methodology: GEC, LF, SJG; Project Administration: XAZ, JG; Resources: XAZ, KTB, JG, DS, SJG; Software: GEC, LF; Supervision: XAZ, PCS, JG; Validation: XAZ, MBB, GEC; Visualization: GEC, LF, XAZ, FLV; Writing - Original Draft Preparation: TML, FLV, MJH, XAZ; Writing - Review and Editing: XAZ, JG, MBB, SJG, GEC, FLV, MJH, TML

Conflict of Interest

The authors state no conflict of interest.